抄録
This paper proposes the method to analyze the musical intervals from recorded musical data by using FFT. We aim to build a system that teaches songs to a robot. Our method was able to reproduce a recorded monotonous song by Beep sound as a result to transform the song into frequency every 50msec. As a result of the experimentation of frequency analysis with the monotonous music and the song of voice synthesis, the small error was measured to sounds of more than 145[Hz].
